Got an error when using 'random_forest_regressor' estimator

[Guo-lingyun]

Since my data had four timepoints-treated conditions, I tried to use 'random_forest_regressor ' estimator which was described for continuous data in the tutorial:
augur = pt.tl.Augur("random_forest_regressor")
treat_augur=augur.load(treat,label_col='group',cell_type_col='celltype')
treat_augur, treat_augur_results = augur.predict(treat_augur, random_state=None, n_threads=4)
However, I got an error in the predicting step:
AttributeError: RandomForestRegressor has none of the following attributes: predict_proba.
What should I do?

[Zethson]

Hi,

is this issue still current?

[jacquelyne-sun88]

Hi there, I have this same exact issue where the regressor is unable to call the function for the continuous variable in 'predict', and continues to call 'predict_proba'. I am on the newest pertpy version and sklearn version.

[Zethson]

Could you please provide a minimal reproducible example @jacquelyne-sun88?

[jacquelyne-sun88]

I’m trying to use Augur in regression mode with a RandomForestRegressor to predict
a continuous phenotype from single-cell expression. However, even when I pass
augur_mode="regression", PertPy/Augur still ends up calling predict_proba
on RandomForestRegressor, which raises:

AttributeError: RandomForestRegressor has none of the following attributes: predict_proba.

Versions:

• pertpy: 1.0.3
• scanpy: 1.11.5
• scikit-learn: 1.9.1
• Python: 3.11.15
• OS: ubuntu

Minimal reproducible example:

my data set is 575584x48228 in an anndata object. and i ran the following codes to perform perturbation analysis using PertPy:
import pertpy as pt
from statsmodels.stats.multitest import multipleAlcohol_Augur.obs['TotalDrinking_gms'] = (pd.to_numeric(Alcohol_Augur.obs['TotalDrinking_gms'], errors='raise').astype('float64'))
ag_rfr = pt.tl.Augur(estimator="random_forest_regressor", random_state=42)
tests
Alcohol_Augur.X = Alcohol_Augur.layers['raw_counts']
loaded_data = ag_rfr.load(Alcohol_Augur, label_col = "TotalDrinking_gms", cell_type_col = "celltypes_gen")
loaded_data
v_adata, v_results = ag_rfr.predict(
loaded_data,
subsample_size=20,
n_threads=12,
select_variance_features=False,
random_state=42,
augur_mode="regression"
)
v_results["summary_metrics"]

The errors are as follows:
Processing data... ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 0% -:--:--
/home/jacquelyne/miniconda3/envs/sc_env/lib/python3.11/site-packages/sklearn/model_selection/_split.py:812: UserWarning: The least populated class in y has only 1 members, which is less than n_splits=3.
warnings.warn(
/home/jacquelyne/miniconda3/envs/sc_env/lib/python3.11/site-packages/sklearn/model_selection/_split.py:812: UserWarning: The least populated class in y has only 1 members, which is less than n_splits=3.
warnings.warn(
/home/jacquelyne/miniconda3/envs/sc_env/lib/python3.11/site-packages/sklearn/model_selection/_split.py:812: UserWarning: The least populated class in y has only 1 members, which is less than n_splits=3.
warnings.warn(
/home/jacquelyne/miniconda3/envs/sc_env/lib/python3.11/site-packages/sklearn/model_selection/_split.py:812: UserWarning: The least populated class in y has only 1 members, which is less than n_splits=3.
warnings.warn(
/home/jacquelyne/miniconda3/envs/sc_env/lib/python3.11/site-packages/sklearn/model_selection/_split.py:812: UserWarning: The least populated class in y has only 1 members, which is less than n_splits=3.
warnings.warn(
Processing data... ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 0% -:--:--
/home/jacquelyne/miniconda3/envs/sc_env/lib/python3.11/site-packages/sklearn/model_selection/_split.py:812: UserWarning: The least populated class in y has only 1 members, which is less than n_splits=3.
warnings.warn(
Processing data... ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 0% -:--:--

AttributeError: RandomForestRegressor has none of the following attributes: predict_proba.

[Zethson]

You're using python 3.11 which hasn't been supported for a while. The latest pertpy version is 1.3.0 not 1.0.3. Could you please upgrade python to 3.14 and install the latest pertpy and try again? If you still run into this issue, I'm happy to have a look.

[Zethson]

@jacquelyne-sun88 this will be fixed with the next release which I'll do this weekend. Thank you for bringing it up again!

In any case, please work with a more recent Python version in the future so that you can use the latest pertpy